Project 1: Co-infection and pathogen shedding in a multi-host, multi-pathogen amphibian disease system

Summary Report

Project Overview

Project Title: “From individual behaviour to population-level transmission: bridging disease ecology scales with the amphibian pathogens.”

Many important pathogens circulate within multi-host communities ². Equally, many hosts can be infected with multiple pathogens at any given time ³. Recent studies of declining amphibian populations indicate that the two pathogens of most concern, Ranavirus and Batrachochytrium dendrobatidis (Bd), commonly occur as co-infections ⁴. Yet, experimental evidence of co-infection dynamics in amphibians remains sparse.

While both pathogens are host generalist, significant heterogeneity exists in host susceptibility, infectiousness and burden of each disease, at a species and individual level ⁵. Which conforms with a wider body of literature showing host species unevenly contribute to disease transmission in multi-host communities ⁶. As a result, pathogen persistence at a community level is largely dictated by the composition of host species ⁷. Predicting how these heterogeneities in disease contribution across different host species relates to the establishment and persistence of a pathogen in a host community, revolves around understanding the transmission process between the infectious stage of a pathogen and a susceptible individual ⁸.

To help unpick the tangled dynamics of this multi-host, multi-pathogen system we conducted an experiment looking at how susceptibility, infectiousness and burden of disease changed with infection scenario (single vs coinfection) across a panel of host species (Bufo bufo, Rana temporaria and Alytes muletensis) that range in their reported susceptibility. We measured the contributions of each host, at an individual level, to the environmental pool of infectious particles, by quantifying Bd zoospores and ranavirus virion outputs for four days post exposure. Endpoint infection load of each host was also measured to link the hosts’ infection burdens with their infectiousness.

Understanding the susceptibility and infectiousness of each host at an individual level allows us to predict how host species community composition influences the establishment and persistence of both pathogens, singularly and as coinfections, at the community level. Our study provides previously lacking emperical evidence of within-host and between-host dynamics under different infection scenarios (single vs co-infection) and highlights the importance of understanding host contributions to enviromental reservoirs of disease.

Aims

I. Ascertain whether the sequence of exposure to co-infecting pathogens (Bd and Rv) alters the disease outcome for the host and subsequent transmission of the pathogens.
II. Establish the contributions of hosts to the environmental “pool” of infectious particles, across species at an individual level, to understand variations in host infectiousness and by proxy transmission.

1. Understand how host infectiousness, measured by quantifying Bd zoospore and ranavirus virons shed into the water body, changes with the exposure regime in single pathogen vs. co-infection scenarios.
2. Assess the viability of infectious particles shed into the aquatic environment.

Methods

Study Species

In order to capture a range of host susceptibilities to infection by Bd and/or Rv I examined shedding rates and infection intensity of two co-occurring species (B. bufo & R. temporaria) that display contrasting resistances, and one other species (A. muletensis) documented to respond distinctly to Bd exposure but which has no empirical records of ranavirus exposure.

Source

B. bufo and R. temporaria clutches were collected, with the permission of the landowner and under licence, from a pond site in Henley as spawn and hatched out in an approved facility. Larval A. muletensis were reared from egg clutches by the ZSL Living Collections.

Numbers

The total number of animals that underwent procedures (includes controls that had sham bath exposures).
Bufo bufo = 220, Rana temporaria = 110, Alytes muletensis = 107

Experiments were run in batches comprising of two B.bufo experiments of 100 animals, with 10 reference animals, a R.temporaria experiment, also with 10 reference animals, one experiment of 70 A.muletensis tadpoles, with all treatment groups, and one Bd-only experiment with less developed A.muletensis (n=37). Each treatment group had n=20 bar the A.muletensis. The larger A.muletensis (here on in referred to as A.muletensis I) had n=14 per treatment and the smaller A.muletensis (A.muletensis II) had n=25 in the Bd-only exposure treatment group and n=12 tadpoles as controls.

The number of tadpoles used in the experiments is based on previous experiments where the lab group have investigated such parameters as endpoint infectious burdens.

Reference animals were euthanised on exposure day 12 (see Figure 1) as a means of assessing whether the exposure schedule resulted in infection. Before euthanasia, the reference animals underwent the shedding/filtering process (see relevant method section below) so we could compare early stage shedding rates with endpoint infection loads and to initially trial the procedure on live samples.

Animal Husbandry

Housing

Tadpoles were transferred to the experimental rooms and into individual housing once the gills were reabsorbed, free swimming and exhibit normal behaviour. For each species, larvae of similar mass and developmental stage (Gosner 25; Gosner, 1960) were randomly selected, acclimatised to the individual housing (Really Useful Boxes (RUBs)) of the appropriate size for the species (see below) and then randomly assigned to a treatment group. As outlined in Figure 1, treatment groups consisted of 20 individuals per species.

All individuals were checked and rotated daily, left to right between treatments and front to back within treatments.

Experiment	RUB size	exposure water vol.	housing water vol.
Bufo bufo I	0.075L	210mL	375mL
Bufo bufo II	0.075L	210mL	375mL
Rana temporaria	0.075L	210mL	375mL
Alytes muletensis I	0.7L	520mL	1000mL
Alytes muletensis II	0.075L	210mL	375mL

Diet

Larval amphibians were fed 200µL of ground Tetra Tabimin tablets dispersed in double-distilled water with a ratio of 1g/100mL, every other day during the experiment. Tadpoles in the first Bufo experiment were fed 300µL for the first few feeds but it was evident they were not consuming all the food so it was reduced down to 220µL. The older A.multenesis (II) tadpoles recieved double the feed (~400µL) due to their larger size.

All the Alytes muletensis received a slightly altered diet. Half of the ground Tetra Tabimin was subsituted with coarse ground Spirulina tablets, in an attempt to improve water quality.

In hindsight I think the volume of food was still too high but unfortunately we didn't record tadpoles weight change so we have no emperical evidence.

Water Changes

Full water changes (100%) were made every 4 days throughout the course of the experiment or before an exposure. A 15ml water samples was collected before the 100% water change pre-exposure and from the hosuing and shedding container after the shedding period for four sentinel tadpoles per experiment.

Water quality was an issue throughout the experimental period. Steps have been taken to resolve this issue for future experiments mainly by the installation of Reverse Osmosis units in the animal rooms - thanks to Chris S.

Temperature

The temperature of the animal rooms followed natural ambient conditions (18-24^oC) and was monitored throughout. The summer heatwave alongside faulty temperate control units meant temperature fluctuated significantly. See Table 2 for minimum and maximum recordings during each experiment. Free standing air-con units were installed as a temporary, emergency measure to control room temperatures by the begining of August 2018. Meaning the temperature for the A.muletensis experiments was carefully controlled.

Experiment	min. temp. (oC)	max. temp. (oC)
Bufo bufo I	16.6	23.5
Bufo bufo II	16.7	27.6
Rana temporaria	16.7	27.6
Alytes muletensis I & II	15	16.6

Table 1. Minimum and maximum temperatures during each experiment measured by thermometer placed in the middle of the animal room. The temperatures for the A.muletensis experiments are readings from a thermometer in the water vat.

Temperature control was an problem for the majority of the experiments. Unfortunately, despite best efforts (thanks again to Chris S.) it looks to be an ongoing issue.

Pathogen Exposure

Tadpoles were exposed individually to controlled doses of Bd and/or Rv or the correlating sham media (see Figure 1). The exposure inoculum was added directly to the housing container after a 100% water change. Following the 6-hour exposure period the water volume in the housing containers was raised to maintain water quality.

Figure 1: Schedule of exposures and sampling, by treatment group. A dose is denoted by a green (Rv) or yellow (Bd) coloured square, and sham doses are displayed as cross-hatched squares. The blue squares indicate when 50ml “soak” water samples were collected from individuals and filtered. EMA water samples (for quantification of viable Bd zoospores) are collected directly from the housing container and shown by an asterisk.

Strain Details

Ranavirus strain: isolate RUK13 ⁹ was cultivated on epithelioma papillosum carp (EPC) cell lines at 18^oC and 5% CO2 (courtesy of S.J.Price, C.Owen and L. Brookes), and quantified using the TCID50 method ¹⁰. The harvested cell culture fluid contained a virus titre of 10⁷ TCID50/ml.

Bd strain: IA’9’13, a member of the hypervirulent BdGPL lineage isolated during an epidemic at Ibón Acherito (Pyrenees, Spain) in 2013 by Prof. M. Fisher, was cultured in TGhL broth, in a 25cm² cell culture flasks, at 18^oC. Zoospores were collected and counted using a haemocytometer.

Dosage Details

Ranavirus Designated individuals were exposed to 105µL (or 263µL for A.muletensis II) meaning an effective exposure of 10^4.5 TCID50/ml. The dose was deemed suitable based on previous work where similar inoculums induced infection but had a longer time till death in tadpoles then higher concentrations ¹¹.

Bd A Bd positive dose consistent of 15,000 to 600,000 active zoospores in 210µL liquid media, or 525µL for A.muletensis II. The volume of media was standardised across doses in order to maintain water quality during the exposure period. Total exposure are shown in the table below.

Experiment	total zoospores
Bufo bufo I	3,675,000
Bufo bufo II	1,443,750
Rana temporaria	2,336,250
Alytes muletensis I	472,500
Alytes muletensis II	294759.9

Within-host Dynamics (co-infection)

Endpoint Tissue Samples

At the end of the experiment (day 9 post-exposure) tadpoles were euthanised, under licence, by buffered (pH 7.0) 5 mg/L tricaine methylsulfonate (MS-222) and then stored in 100% ethanol.

The Bd infection load of tadpoles was confirmed by excision and DNA extraction of mouthparts, the site of Bd colonization and infection in tadpoles, using Prepman Ultra (Life Technologies) as per Hyatt et al. (2007). Extracts were screened by the qPCR diagnostic outlined in Boyle et al. (2004) which targets the ITS-1 and 5.8S regions and diluted 1:10 before qPCR to avoid inhibition. The infection load with be express in genomic equivalents (GE), where one GE is equivalent to a single zoospore.

The quantification of ranaviral DNA from tissue samples (mainly livers and kidneys) of the tadpoles stored in 100% ethanol was by DNeasy Blood and Tissue (Qiagen) extraction following the manufacturers protocol. DNA samples were then analysed with a qPCR assay specific to the ranaviral major capsid protein (MCP) sequence and normalised by host cell quantity as outlined in Leung et al. (2017).

Figure 2: Left DNA extraction procedure for endpoint tissue samples. Right

Between-host Dynamics (shedding)

Shedding Samples

One limitation of using common frog and common toad larvae is that individuals can not be sampled for infections without first being euthanized. For this reason, we tracked infectious particle output post-exposure using a modified version of the “soak” technique ¹². Individuals were transferred into a temporary shedding unit containing clean, aged water to “soak” for 4 hours, after which they will be returned to their housing unit. The shedding containers used are shown in Figure 3. The soak period of 4 hours was choosen to balance the need to sensitively detect zoospores/virions shed with the considerations of the tadpoles welfare needs.

Figure 3: Images of the disposable containers used as temporary holding units to generate a water sample of shed pathogens from the tadpoles. Left shows the standard shedding container for all experiments,, where the water volume was 50mL, right shows the shedding containers used for the larger Alytes muletensis tadpoles, where the water volume was 100mL.

The soak water was filtered through a cellulose nitrate filter membrane (Nalgene Analytical Test Filter Funnel, ThermoFisher) with 0.45µm pore size to capture 1-2µm to 3-5µm zoospores ¹³, by a vacuum manifold, see Figure 4. Following filtration, each membrane was removed from the filter unit, cut in half using a sterile scalpel blade, and stored at -20^oC for a week before being transferred to -80^oC until processing. DNA extractions were performed for each filter membrane half. One half underwent extraction by Prepman Ultra (Life Technologies) following the procedure described in Hyatt et al. (2007) to target Bd. Whereas DNeasy Blood & Tissue kit was used to isolate Rv DNA following the protocol Goldberg et al.(2011) with the modifications outlined by Kosch & Summers (2013). All DNA extractions were assayed by quantitative polymerase chain reaction (qPCR) specific to each pathogen and run in duplicate. A positive result was scored if both wells amplify above the detection threshold when compared to the curve of standards.

Figure 4: Setup of the vacuum manifold with the cellulose nitrate filter units.

Water samples were obtained every day for the first four days and then day 9 post-exposure, to capture the Bd reproduction cycle ¹⁴ and Rv attenuation ¹⁵. See Figure 1) for sampling schedule. After the last shedding sample (day 9 post-exposure) tadpoles were immediately euthanised, under licence, by buffered (pH 7.0) 5 mg/L tricaine methylsulfonate (MS- 222) and then stored in 100% ethanol.

Viability (EMA) Samples

Quantification of viable zoospores shed into the housing container over the course of the shedding sampling period (4 days) was achieved by ethidium monoazide (EMA) treatment following the protocol in Blooi et al. (2013), see Figure 5 for an overview. In brief, at 4 days post-infection, two sub-samples were taken from the housing container, one to be treated with EMA and the other untreated before DNA extraction by Prepman Ultra. EMA binds to dead zoospores penetrating their compromised membranes and blocks qPCR amplification thus EMA treated samples represent the viable proportion of Bd zoospores in a sample compared to an untreated sample that records both the viable and dead fractions.

Figure 5: Viability staining protocol for quantification of live vs. dead Bd zoospores.

Results

Mortality

We saw some mortality across all species and treatment groups. In the Bufo bufo we saw increase mortality in the coinfection treatment group (Rv>Bd). Several individuals in this treatment group showed signs of disease. The main signs observed were ragged tail fin, signs of haemorraghages (most often on the tail) and edemas.

Survival Analysis

results pending

Within-host Dynamics (co-infection)

Endpoint Infection Status

Figure 6: Proportion of individuals infected, by pathogen, within a treatment group for the three host species. N.B. that Alytes II is excluded.

Endpoint Infection Load

Bd endpoint pathogen load (see Figure 7)

Figure 7: Boxplot of the endpoint Bd load for each treatment group across the three host species. Bd load is quantified as genomic equivalents (GE) where 1 GE represents 1 Bd zoospore. The black dots represent each sample.

Ranavirus endpoint pathogen load (see Figure 8)

Figure 8: Boxplot of the endpoint Ranaviral load for each treatment group across the three host species. Viral load has been normalised using Leung et al.’s (2017) method. The black dots represent each sample.

Between-host Dynamics (Shedding)

results pending

Future Project Direction

Eventually, I hope to use my experimental outputs to construct mathematical models that realistically replicate the disease dynamics of the pathogens, which will enable me to test how pathogen persistence, amplification and/or loss might occur under different scenarios of host community complexity and coinfections.

Future Aims

1. Build an understanding of the spatial dynamics of Bd zoospores in the aquatic environment, focussing on zoospore activity and trajectory.
   
2. Parametrizing the transmission coefficients of Bd in a multi-host system to develop realistic models.
   
3. Combine these results into a predictive framework to understand how individual-level behaviours influence disease transmission in natural ecological communities.
   
4. Elucidate the role of specific behaviours in modifying contact rates and how that influences transmission rates, within and between species.
   
   1. Establish whether exposure to pathogens (Bd and/or Rv) changes tadpole behaviour (e.g. activity rates, foraging performance, and aggregation behaviour).
      
   2. Identify which species-to-species and individual-to-individual contacts alter pathogen transmission.
      
   3. Assess how transmission rates change under different scenarios that encourage shifts in contact rates (e.g. variations in host density, temperature and food availability).
      

Feedback

Please email any comments to bryony.allen@liverpool.ac.uk. Here are the links to my website bio.

Please note all graphics were created with BioRender.

References

---

1. University of Liverpool & Zoological Society of London↩

2. Holt et al., 2003; Webster, Borlase, & Rudge, 2017; Woolhouse, Taylor, & Haydon, 2001↩

3. Hellard, Fouchet, Vavre, & Pontier, 2015; Petney & Andrews, 1998; Rigaud, Perrot-Minnot, & Brown, 2010↩

4. Reshetnikov et al., 2014; Rosa et al., 2017; Souza et al., 2012↩

5. DiRenzo, Langhammer, Zamudio, & Lips, 2014a; Fernández-Beaskoetxea et al., 2016; Reeder, Pessier, Vredenburg, & Litvintseva, 2012; Searle et al., 2011; Gray & Chinchar, 2015; Schock, Bollinger, Gregory Chinchar, Jancovich, & Collins, 2008↩

6. Begon et al., 1999; Duffus, Nichols, & Garner, 2014; Fenton et al., 2015; Fernández-Beaskoetxea, Bosch, & Bielby, 2016; Woolhouse et al., 1997↩

7. Holt et al., 2003; Keesing, Holt, & Ostfeld, 2006↩

8. Begon et al., 2002; McCallum, Barlow, & Hone, 2001; McCallum et al., 2017↩

9. Cunningham, Hyatt, Russell, & Bennett, 2007↩

10. Reed & Muench, 1938↩

11. Duffus et al., 2014; Pearman, Garner, Straub, & Greber, 2004↩

12. Di Renzo et al. (2014); Reeder et al. (2012)↩

13. Berger, Speare, & Kent, 1999; Longcore et al., 1999↩

14. DiRenzo, Langhammer, Zamudio, & Lips, 2014b; Garmyn et al., 2012↩

15. Duffus et al., 2014↩